Vector Quantizer design by conjugate gradient optimized hyperplane

An Vector Quantizer design method by adaptive hyperplane generation using conjugate gradient optimization is proposed. The generated hyperplane is a perpendicular bisector of the clustering set centroids at each stage of the K-dimensional search tree, thus eliminated misclassification error associated with hyperplane based vector quantization. Simulation results on Vector quantization image coding is presented and compared with that obtained by other algorithms in literature. Where the results showed that the proposed algorithm can achieve better PSNR image coding results than that obtained by other algorithms. The generated K-dimensional search tree vector quantizer facilities computational efficient quantization process.